Integrative Analysis of Identifying Methylation-Driven Genes Signature Predicts Prognosis in Colorectal Carcinoma

Abstract

BackgroundAberrant DNA methylation is a critical regulator of gene expression and plays a crucial role in the occurrence, progression, and prognosis of colorectal cancer (CRC). We aimed to identify methylation-driven genes by integrative epigenetic and transcriptomic analysis to predict the prognosis of CRC patients.MethodsMethylation-driven genes were selected for CRC using a MethylMix algorithm and LASSO regression screening strategy, and were further used to construct a prognostic risk-assessment model. The Cancer Genome Atlas (TCGA) database was obtained as the training set for both the screening of methylation-driven genes and the effect of genes signature on CRC prognosis. Then, the prognostic genes signature was validated in three independent expression arrays of CRC data from Gene Expression Omnibus (GEO).ResultsWe identified 143 methylation-driven genes, of which the combination of BATF, PHYHIPL, RBP1, and PNPLA4 expression levels was screened as a better prognostic model with the best area under the curve (AUC) (AUC = 0.876). Compared with patients in the low-risk group, CRC patients in the high-risk group had significantly poorer overall survival in the training set (HR = 2.184, 95% CI: 1.404–3.396, P < 0.001). Similar results were observed in the validation set. Moreover, VanderWeele’s mediation analysis indicated that the effect of methylation on prognosis was mediated by the levels of their expression (HRindirect = 1.473, P = 0.001, Proportion mediated, 69.10%).ConclusionsWe identified a four-gene prognostic signature by integrative analysis and developed a risk-assessment model that is significantly associated with patients’ survival. Methylation-driven genes might be a potential prognostic signature for CRC patients.

Machine learning in the identification of prognostic DNA methylation biomarkers among patients with cancer: a systematic review of epigenome-wide studies

“…In the stage of model training, further feature selection was performed by more than half of the 30 studies (8, 9, 24, 25, 28, 31, 36, 40-42, 48, 51, 66, 68, 72-74, 81, 84, 85). Sample-splitting was the most commonly used internal model validation method ( N = 12), and external validation was performed in 15 studies (7, 8, 24-26, 31, 42, 48, 59, 66-68, 72, 74, 81).…”

“…Similar to clustering-based workflows, an initial screening was generally the first step for the 30 studies using supervised feature selection-based workflows (Figure 4.2) (6-9, 24-28, 30, 31, 36, 40-42, 48, 51, 59, 60, 63, 66-68, 71-74, 81, 84, 85). In the next step, nearly all studies used one to three types of supervised feature selection techniques to pre-select prognostically relevant DNA methylation biomarkers prior to training one or more prognostic models.…”

“…Fewer studies used other metrics, including C-index ( N =16) and receiver operating characteristic curves ( N = 39) to evaluate the discriminative power of their prognostic models. Only 18 studies (6, 31, 36, 39-42, 51, 54, 55, 57, 59, 66, 67, 78, 80, 84, 88) evaluated the calibration accuracy of prognostic models.…”

“…Lastly, we found that 18 identified studies first constructed a Cox prognostic model based on single selected CpG sites (i.e., methylation score), which was subsequently used as a single predictor, together with other clinical variables, to construct a second Cox prognostic model (6, 31, 36, 39-42, 55, 57, 66-69, 73, 78, 80, 81). It might be an inappropriate approach to construct the first “methylation score”, in which the relative weight of individual CpGs were consequently fixed for the second prognostic model.…”

Comprehensive Analysis of DNA Methylation and Transcriptome to Identify PD-1-Negative Prognostic Methylated Signature in Endometrial Carcinoma

“…In high-grade serous ovarian cancer, hypermethylated NCALD and LAMA3 are associated with prognosis and chemotherapy resistance in advanced patients [ 12 ]. In addition, a number of studies have identified panels of multiple methylation driving genes to predict the prognosis of colorectal cancer [ 13 ], gastric cancer [ 14 ], and hepatocellular carcinoma [ 15 ]. However, there is a lack of a comprehensive panel of methylation markers for EC [ 16 ].…”